Loading...
×Use a Paperclip reporting agent with Tareno configured in its Claude Code or Hermes runtime to turn closed issues into an X progress post that names its evidence, saved as a draft you approve.
Yes, once Tareno is configured in the agent's runtime, because Paperclip has no MCP command of its own. A reporting agent reads the closed issues, writes a post within 280 characters that names them and saves a draft through Tareno. Tareno has not tested this end to end, and a time request stays pending until you approve it.
Last reviewed 2026-10-25 · Markdown version for AI agents
A company of agents produces plenty of work and very little of it is told to anyone. Paperclip keeps the record: issues, runs, budgets. A reporting agent hired for that purpose can read the closed issues of a goal and describe the result, which is a better source for a progress post than a human's memory of the week. Paperclip does not add MCP servers itself, so Tareno is configured in the runtime that the agent uses, and each heartbeat inherits it. This page follows an operator who runs a small research company of agents and posts a weekly update on X. It keeps the unfinished work out of the post and the human as the only one who can publish. Tareno has not tested this end to end.
Imagine an operator whose Paperclip company works on a literature review. Tareno was added once in the runtime her reporting agent uses, and she completed the sign-in as the same operating-system user that runs the heartbeat. On Friday the agent reads the goal's closed issues, finds that forty papers were screened and eight summarised, and writes a post within 280 characters with issue numbers listed beneath it. She checks two of them. The agent confirms her X account with list_accounts and saves the text with create_drafts. She picks a time in the Tareno dashboard.
Follow the Paperclip overview: add Tareno in the Claude Code adapter's user scope or in Hermes's config, and complete the OAuth flow once as the user that runs the heartbeat. Paperclip itself has no command for this.
Check the server from the agent's working directory with the environment the heartbeat uses, then run a read-only list_accounts in a run. The run viewer shows which tools were available to the agent.
Give it one goal and one week. It should list closed issues with numbers and pick the single result a stranger would find interesting, skipping work that is blocked or still running.
Pick two or three issues the post names and open them. An issue closed as won't fix or duplicate is not a result, so the count in the post must exclude it before you go further.
The agent checks the account via list_accounts and files the approved text through create_drafts, returning a draft ID. No time exists yet.
Give a day, a time and an IANA zone. schedule_posts leaves a pending action in the dashboard for you to release.
Use Tareno to list my connected social accounts with platform and account name. Read only: do not create, schedule or publish anything. If you cannot reach Tareno from this run, say so clearly and report which tools you did have available.Read the issues closed for goal [literature review] in the last seven days. List the issue numbers, mark any closed as duplicate or won't fix, and name the one finished result an outsider would care about. Ignore anything still blocked, and do not write the post yet.Write an X post within 280 characters about that result, with the number of papers screened and summarised, no hashtags and no internal ticket links. Under the post, list the issue numbers each figure rests on so that I can check them.Use Tareno to confirm my X account [handle] and save the approved post as a draft. Return the draft ID and the text, and request no time until I give you the day and the time zone.Take the draft [draft ID] and ask Tareno to schedule it for [Friday 16:00] in [Europe/London]. Return the action ID and remind me that the request still waits for my approval in the Tareno dashboard.As configured in Tareno on 2026-10-24. The live rules for your own account come from the get_platform_schema tool.
Source: X API documentation: create a post, read on 2026-10-24. Platform rules change; check the page before you rely on a number.
https://tareno.co/api/mcpUse the same Tareno account across supported clients. Open the relevant setup guide for authentication, first use, reconnect, and revocation.
list_accountsChecked:
In the runtime the agent uses, not in Paperclip. Paperclip's guide documents Claude Code and Hermes adapters, with user scope or Hermes's host-wide YAML for remote servers. The control plane launches the runtime and only schedules the heartbeats.
Paperclip's guide says to complete the OAuth flow as the same operating-system user that runs the heartbeat. Tokens are stored per user, so another user's heartbeat will not find them and the tool list will come up empty.
The agent proposes the result that looks most finished, and you decide whether the company wants to say it in public. Always read the draft in Tareno before any time is requested.
Run the server check from the agent's directory, then a read-only `list_accounts` inside a heartbeat. Paperclip's guide and Tareno's own setups support this, but no one has run the whole chain end to end.
No. Each time request is a pending action in the Tareno dashboard, and nothing reaches X before you approve it. Afterwards `get_action_status` reports what X did with it.
Only the agent's own run costs do. A weekly reporter with small runs stays within the limit, and Tareno applies no such budget; its protection is the approval step before anything is published.